INVERSE ANALYSIS OF GEOTECHNIC PARAMETERS ON IMPROVED SOFT BANGKOK CLAY by D.T. Bergado, A.S. Enriquez, C.L. Sampaco, M.C. Alfaro and A.S. Balasubramaniam Reprinted from Journal of Geotechnical Engineering, Vol.118, No.7 1992, pp.1012~1030 271INVERSE ANALYSIS OF GEOTECHNIC PARAMETERS ON IMPROVED Sorr BANGKOK CLAY By,Dennes T. Bergado,! Associate Member, ASCE, Apollo S. Enriquez,? Casan L. Sampaco,? Marolo-C. Alfaro,* and A. S, Balasubramaniam,’ Fellow, ASCE Aestract: Deformatisn, strength, and flow parameters were determined by iri= verse analyses: cotechi data obtained from two full-scale test embankments constructed coved soft Bangkok clay. One site was improved with pre ical drains (PVD), and an adjacent site was improved with compacted granular piles (CGP). In the PVD site, the ratio of c, (field) to cy (lab) ranged fcom 1.5~4, and the cy value was 4.90 myear. The value of ky/k, was about 10, - with dd, of 2.5. In the CGP site, the ratio of ¢, (field) to ¢, (lub) ranged from 1,3-12. The smeared zone was considered by reduction to CGP diameter by one- .. fifth. The ratio of E, 10 $,, was determined to be 150‘The observed final settlements ‘in the CGP site was 30-35% lower than those in the PVD site. Using back-analyzed soil parameters, the settlemerits of the test embankments were predicted well. Both ground improvement schemes are effective in improving the soft Bangkok clay. INTRODUCTION Recently, back or inverse analyses in both laboratory models and field prototypes have indicated promising results, and they have become an ac- model of validating theories and evaluating soil parameters for use in available and accepted methods. The accuracy with which an. inverse analysis is carried out depends on the working assumptions and. knowledge ‘of the advantages and limitations of the theoretical models being.used and validated. This study aims to evaluate the deformation, strength, and flow parameters of the soft Bangkok clay improved by prefabricated vertical drains (PVD) and compacted granular piles (CGP) in conjunction with enibankment preloading, by back calculation of soil parameters using avail- able field data. The detailed description of the improvement and preliminary data were reported elsewhere (Bergado et al. 1988a, b). ‘Site Location AND Soi PROFILE The test site is on the campus of the Asian Institute of Technology, about 42 km north of Bangkok on the central plain (Chao Phraya) of Thailand. The subsoil profile reveals the three topmost sublayers of Bangkok marine clay deposits underlain by alternating layers of very dense sand and stiff ‘Assoc. Prof., GTE Div., Asian Inst. of Tech., G.P.O. Box 2754, Bangkok 10501, Thailand. Soil Engr., Moh and Assoc.; Inc,, Taipei, Taiwan, Republic of China. Res. Assoc., GTE Div., Asian Inst. of Tech., G.P.O. Box 2754, Bangkok 10501, ‘Vhailand. ses. Assoc., GTE Div., Asian Inst. of Tech., G.P.0, Box 2754, Bangkok 10501, Thailand, SProf., GTE Div., Asian Inst. of Tech., G.P.O. Box 2754, Bangkok 10501, Thai- land. Note. Discussion open until December 1, 1992. To extend the closing date one month, a written request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on May 8, 1990. This paper is part of the Journal of Geotechnical Engineering, Vol. 118, No. 7, July, 1992, ©ASCE, ISSN 0733-9410/92/0007-1012/$1.00 + $.15 per page. Paper No. 26640.clay extending to dep.» of more than 500 m. The Bangkok clay at the site consists of the uppermost 14-15 m, which can be divided into three distinct layers: namely, weathered clay, soft clay, and stiff clay (Fig. 1). The depth of the water table varies seasonally from 0.5-2.5 m, averaging about 1.5 m. The index properties of the compressible soft clay are shown in Fig. 1. The stress history at the site is depicted in Fig. 2. Field vane tests were carried out using a Geonor vane apparatus at depths of 8 to 9 m. Typical results also are shown in Fig. 1. GROUND IMPROVEMENT BY PREFABRICATED VERTICAL DRAINS Barron (1948) presented a solution to the differential equation for prob- Jem of consolidation of a soil cylinder containing a central drain. The dif- ferential equation formulated in terms of polar coordinates is expressed as follows: au lou | au —= ieee mer | el eet leat stots tee tate ert 1 a (: ar * =) % where u = the excess pore pressure at any point and at any time by virtue of radial flow only; r = the radial distance of the considered point.from the center of the drained soil cylinder; ¢ = the time after an instantaneous increase of the total vertical stress; and c, = the horizontal coefficient of consolidation, From (1), the general expression for average degree of con- solidation due to horizontal drainage derived by Barron (1948) was modified by Hansbo (1979) as follows: 7 —8T, h Os XD | eet eee ere en eerie cert (2) F where ATTERGEAG LIMITS ANO | UNIT WEIGHT | VANE SHEAR STRENGTH Som Lavers — |NATURALWATER CONTENT(%=)] —— (AN/m?) (010 kN WD z0 49 6 80 Wo | § 6 IP 20.30 40 ‘Routt 10 REDOISH 1 -] BROWN WeaTHERED cua 2 ono 34 o_o 4 o—+0 & © Tank crar sort ctay, ag 5-forten ware vecom— 6 # * feos wooo ano & , [sanov seams ge e oo 7 — 24 o4 © Test NOt tofticur Grey sree [PL Wa Lt a TEST NO 2 clay oo © TEST 10.3 FIG. 1. Subsoll Profile at Site with Corresponding Index Properties and Undrained Shear StrengthEa = a O—O Finan vornical sTness = (uewuane, 1842) & \ A rina verricaL sess Bos , (roulos anus 1978) MAX PAST PRESSURE (rms stuor $ 2,0,0 MAK. PAST PRESSURE, é k (9e3-i9aah “0 8 TRO fn 4B 200 AO. STRESS (KN/ qm) FIG. 2, Stress History at Site with Results of Previous Investigations and D, = the diameter of the equivalent soil cylinder; F = the factor that expresses the additive effect due to the spacing of drains F(n), the smear effect F,, and well resistance: F,.Hansbo (1979) gave the following expres- sions for the. spacing, well resistance, and smear. factors,-respectively: F(n) = In (2) - 3 ae a ee oe ee (4) = — 2) (% F, = m2(L - 2) (#) bo ee ee bvbeveeseeeeeetevenswaeedia weve (8) = hey d, : iafres)u(S) coos ome “oie Ri) eth ee eee where d, = the diameter of the smeared zone; L = the length of the drain when opened at one end only; z =; thé vertical distance fromopen end of drain; k, = the permeability of the disturbed zoné in the horizontal direc- tion; and q,, = the discharge capacity of the drain. Recently, FEM analysis (Rixner et al. 1986) showed that the equivalent diameter of the drain (d,) preferable for use ‘in practice can be obtained aswhere a and 6 = the thickness and width of the band-shaped drain, re- spectively. The combination of vertical and radial flow was presented by Carillo (1942) inthe following form: DL (1— OL Oy) reece gee eee levee eee ngneeeeeeeees (9) where U = the combined average degree of consolidation due to vertical and horizontal drainagé: U, = the average degree’ of consolidation due to vertical drainage; and U, = the average degree of consolidation due to horizontal drainage. : GROUND IMPROVEMENT UsING CoMPACTED GRANULAR PILES From the unit cell concept, the distribution of vertical stress within a unit cell can be expressed by an area replacement ratio a,, which is defined as the ratio of the cross-sectional area of the granular pile to the total composite area of the unit cell. The average loading intensity o can be calculated as o = 0,4, + o,(1 - a) Introducing the stress concentration factor n, which is defined as the ratio of the stress in the granular pile, o,, to the stress in the surrounding cohesive soil, o, the stresses in the clay and granular pile are, respectively, given as a a mere le ee) eet ttre L "TE Da, * (uy) ng Oe Gea, ME critter tien (12) The equilibrium method (Aboshi et al.’ 1979) of estimating the settlement of composite ground, which was based -on conventional, one-dimensional consolidation theory, was found to be Jéss realistic and theoretically sound than the finite element method. The finite element method incorporates material nonlinearity, interface slip, and suitable boundary-conditions. So- lutions in terms of nonlinear finite element design. curves. to estimate the settlement of the composite ground were presented by Balaam et al. (1977). Procepures A 4.0-m-high test embankment preloading was constructed on improved ground with prefabricated vertical (PV) drains to study the effectiveness of PV drains on soft Bangkok clay. The rate of embankment loading is shown in Fig. 3. The drains were installed in triangular pattern at 1.5-mt center-to- center spacing by means of a special mandrel to a depth of 8,0 m. A rec- tangular-shaped mandrel was used, which had inner dimensions of 28:0 mm by 133.0. mm and outer dimensions of,45.0 mm by 150,0 mm, just large enough to, contain the 3.0-mm by,95.0-mm Mebra prefabricated vertical drains made of geotextiles. The embankment plan and section views, in- cinding the layout of prefabricated vertical drains, are shown in Figs. 4(a) and (b). A 2.4-m-high test embankment was constructed on improved ground withgranular piles. After 345 days, the test embankment was raised to a height of 4.0 m to provide a meaningful basis of comparison with the performance of a nearby test embankment constructed on improved ground using pre- fabricated vertical drains (Fig. 5). The granular piles were 300.0 mm in diameter and 8 m long, fully penetrating the soft clay layer. The granular piles were arranged in a triangular pattern with a center-to-center spacing of 1.5 m. They were installed by first driving an open-ended steel casing. A large auger was then used to remove. the clay inside the casing. The granular materials, consisting of 20-mm maximum size sandy gravel, were carefully added in layers and compacted by means of dropping a 1.6-kN hammer at a height of 0.6m to a steel disc placed on the surface of the granular material as the casing was withdrawn. Fifteen blows were applied to each layer, compacting the thickness of each layer by about.0.60 m. The friction angles of compacted gravel were obtained from direct shear tests using a shear box dimension of 100 nim x 100 mm x 60mm, and they were found to vary from.39—45°, with compacted densities ranging from 17-18.1 kN/m>. The plan and cross section of the embankment, including the layout of field instrumentation, are illustrated in Figs. 6(a) and (b). The solution suggested by Olson (1977) for time-dependent loading was not employed, because the rate of loading was fast enough to be considered instantaneous, as demonstrated in Figs. 3 and 5. TESTS AND INSTRUMENTATIONS = The locations of sutface and subsurface settlement gauges, piezometers, and inclinometers for,PVD and CGP sites are shown in Figs. 4(a) and (b) and 6(a) and (b), respectively. Additional field tests and sampling were made and additional monitoring instruments are installed, as described in the following. Three boreholes were drilled down to a 9.0-m depth. Two of these boreholes (BH1 and BH2) were made through the 4.0-m-high embankments on compacted granular piles and prefabricated vertical draifis, respectively. The third borehole (BH3) was drilled on natural unimproved ground. A 3-in.-diameter (76.0- LOADING HISTORY EMBANKMENT HEIGHT (m) ° 200 400 600 800 1000 1200 TIME {DAYS} FIG. 5. Embankment Loading History at CGP SiteLORDING HISTORY EMBANKMENT. HEIGHT (m) ° 200 © 400 600 800 wooo "1200 TIME (DAYS) - FIG.°3. ‘Embankment Loading History at PVD Site FIG. 4, Embankment Plan, Section View, and Layout of Fleld Instrumentation at PVD Site(a) (64 FIG. 6. Plan and Embankment Cross-Section Including Layout of Field Instru- mentation at CGP Site ° mm) thin-walled tube piston sampler was used, and continuous sampling was done by pushing manually. Fourteen conventional consolidation tests were conducted, consisting of six sets for samples under each of the two embankments and two sets for the unimproved samples. In the horizontal oedometer tests, samples were cut at orientations perpendicular to their original directions in the field, as suggested by Laminen and Rathmeyer (1981). A total of six field vane shear tests were also conducted to depths of 8-m at the site: two each on the subsoil under both embankments and two on natural unimproved ground. Aside from the original eight, previously installed, closed hydraulic piezometers, an additional seven open standpipe piezometers were installed to monitor the variation of pore pressures within and outside of the test embankments: Two were installed in each embank- ment of 3.0-m and 6.0-m depths. The other three were installed 10 m away at the same depths as dummy piezometers. The fluctuations of ground-water levels were monitored by means of an observation well consisting of per- forated pipes. Inclinometers had been installed 12.0 m deep for the PVD site and 13.0 m deep for the CGP site. to measure subsurface lateral dis- placements. There were three surface settlement plates installed in eachembankment. For the CGP site, two were installed on top of the granular piles and one was placed on natural ground. For subsurface settlements, four settlement plates were originally installed below the CGP site at 1.5- m, 3.0-m, 6.0-m, and 8.0-m depths, while three plates were installed below the PYD site at 2.5-m, 5.5-m, and 7.0-m depths. The bench mark was a deep well casing installed 200 m below ground level. The embankment on compacted granular piles was instrumented with two earth pressure cells; one was placed on top of one of the piles and the other was placed on the original ground surface between the granular piles. ANALYSIS OF LABORATORY AND FiELD DATA Coefficients of Consolidation and Compression Index The laboratory c,, and c, values were obtained using the square-root-of- time fitting method and were found to decrease significantly as the consol- idation stress approached the preconsolidation pressures. They then re- mained approximately constant in the normally consolidated region. Typical results are illustrated in Figs. 7(a) and (b) for vertical and horizontal oed- ometer tests on samples from PVD sites, respectively. The selected values of c, and c, are those that correspond to the final effective vertical stresses anticipated in the field. In most cases, the coefficient of consolidation for both horizontal and vertical directions showed Jess variation in the normally consolidated region. For CGP site, the range obtained for (¢,/c,.) is about 1-4. For PVD site, the corresponding ratio ranged from 1-7. The variation of the compression index is given in Fig. 8. The soft clay layer between the depths of 2.5-5.0 m is the most compressible zone, as shown in this figure. a pvp sag. aopooe (a) toe oF. Pressure 7m") (b) / Woe oF mmesiunt) Oat FIG. 7. Values of c, and ¢, on Samples Obtained from PVD SiteDEPTH (mY TEST NO. a TEST NO. 2 TEST NO. 3 on os. os OF 08. tH us us COMPRESSION: INDEX, “Cy FIG. 8, Variations of Compression Index with Depth STRESS. CONCENTRATION FAETOR (n} ° 200) ‘400 600 00) 1000 1200 “TIMES (DAYS) FIG. 9, Measured Stress Concentration Factor (1) with Time Stress Concentration Factor (n) Set The values of the stress concentration factor, n, as measured from the two total earth pressure cells installed on top.of the granular pile and in between the piles, are shown in Fig. 9. They initially increased at the initial stage of loading. This initial increase is partially due to the passive earth resistance of the soil, which is progressively mobilized to restrain the lateral dilatancy of the granular piles. As the surrounding soil consolidated, the load on the piles increased. In this study, low area replacement ratio, a,, was used:and the measured value of n decreased to 1.34 (Fig. 9). This result is contrary to that reported in the literature (Juran and Guermazi 1988), in which the value of the stress concentration factor increased with time (i.e., when the surrounding soil consolidated). The possible reasons for these variations in observation include lateral yictding of the piles resulting from the load transfer from the piles, to the surrounding soil, the effects of low replacement ratio, and inaccuracy of the total earth pressure cells due to arching effects caused by the. presence of stich instruments.SETTLEMENT ANALYSIS Fig. 10 shows the measured surface settlements at the CGP site. At the initial embankment height of 2.40 m, the maximum measured surface set- tlement was only about 8:80 cm. The small recorded settlements may be due to the effect of stress concentration on the stiffer piles.and the low applied load, such that the increase in vertical stress, applied_to the foun- dation subsoil from this loading, was not enough to exceed the existing preconsolidation pressures of the foundation. An increase in:the rate and magnitude of settlement was observed when the embankment height was increased to 4.0 m. The surface settlement plate $2 on natural ground registered a little higher magnitude than settlement plates S1, and S3, which were on top of two granular piles. The difference is minimal, this the “equal strain theory” (Barron 1948) is considered applicable and acceptable. The observed subsurface settlements are shown in Fig. 11. As expected, much higher settlements occurred at shallower depths. cop STE. SURFACE _ SETTLEMENTS 81 (ON STONE) 52 (ON GROUND } 3 (ON STONE) SETTLEMENT (em) MARCH 20, 1989 mwe (0ayS) FIG. 10. Measured Surface Settlements at CGP Site We]. cop. site va]. SUBSURFACE. SETTLEM SETTLEMENT (em) 14} 0 sb 1 tism) © S03 (3.0m) 20, © S82 (6.0m), 4 $84 (8.0m) MARCH 20, 1989, HME ( oars} FIG. -11., Measured Subsurface Settlements at CGP. SiteFig. 12 shows the. measured surface settlements at the PVD site. Higher settlements occurred in $2, which was at the center of theeembankment. Fig. 13 shows the measured subsurface settlements at the PVD site. The maximum subsurface settlement. obtained, which included some effect of secondary compression, was about 292.0 mm after 1,200 days. This was derived from SS1, which was near the center of the embankment ata shallow depth of 2.50 m below the ground surface. Asaoka (1978) suggested a graphical method to determine the final set- tlement. The observed time-settlement curve plotted to an arithmetic scale was divided into equal time intervals, At (usually between 10 and 100 days). The settlements p,, p2,... , corresponding to times 4, ,°... . , are read off and plotted as points (p,, py-;) in the coordinate system, as.shown in Figs. 14:and 15 for PVD and cop sites, respectively. A straight line is fitted through the points. The slope of this line is B,, and its intercept with the ordinate axis is By. The 45° line with py = py-, is also plotted. The point where the line intersects the 45° line gives the final settlement, py. “pvp site SURFACE SETTLEMENTS ost © sz SETTLEMENT em) MARCH 20,1989 ° 200 400 600-800 1000 1200 2 TIME. (DAYS) FIG. 12... Measured Surface Settiements at PVD Site Pvo SITE REA MENTS D ssi (2.50m) £ © $82 (5.0m) . 5 i a g E 3 8 g i ° 3,200 - 400 eo 600 1000 +) 00, TIME, (DAYS) IG. 13. Measured Subsurface Settlements at PVD Site _“ashoke's” OBSERVATIONAL METHOD “agnoxa's” oaseRviioniAl MET HOO (Pv SITE) ~~ (GP sive) , at grace [CoNSOLBATION| BRiMARY CONSOLIDATION Heyer Cy [ WITH cy or Cy DECREASING) (CONSTANT), f+ SLOPE OF FITTED tine As NTERCEPT OF "FITTED LINE he lgcren] A | certs | teh 25_ Jo 9082 ].375] 36 00] 0 5 6 207% 30 38 Py fom) P,_yfem) FIG. 14. Determination of Final Settle- FIG. 15. Determination of Final Settle- ment Using Asaoka’s. Graphical Con- ment Using Asaoka's Graphical Con- struction (PVD Site) struction (CGP Site) The final settlement at the CGP site calculated from Asaoka:(1978) is about 250.0 mm. The settlement at 90% degree of consolidation occurred at about 880 days from the start of construction. The-final settlement without granular piles is 360.0 mm, as derived from the settlement records of the PVD site using Asaoka (1978). This would mean a settlement reduction of about 30-35% due to the presence of the CGP. Primary consolidation was estimated to end at 1,040 days after the beginning of construction. The final settlement in the PVD site is about 360.0 mm. This final settlement occurred 610 days after the start of construction. The settlement at 90% consolidation was observed 430. days from the beginning of construction. Comparison of the surface settlements for both sites is shown in Fig. 16. There was 30-35% reduction of settlement by using. granular pil For the case of subsurface settlements, the comparison for both embankments is depicted in Fig. 17. It is interesting to note that the settlement at the 5.50- m depth for the PVD site is much greater than settlement at.the surface of the CGP site. This could be the result of the stress concentration and rein- forcing effects of granular piles. INVERSE ANALYSIS. The soil parameters determined from available field data were the coef- ficients of consolidation, ¢,,and c,, and undrained modulus, E,. In the case of the drained condition, the Poisson ratio was based from previous studies with values of v’ = 0,26 for the weathered*clay and v' = 0.24 for the soft clay layer. The evaluation of drained settlements was analyzed using Asaoka (1978), which made possible’a separate determination of primary consoli- dation settlement and the soil consolidation rate on the basis of settlement- versus-time records. When applied to consolidation problems with the pres- ence of vertical drains, Asaoka (1978) is based on the solution of Barron (1948) for purely radial drainage. Asa result, some modifications were made in this study to introduce Hansbo’s equation (Hansbo 1979), which considers the smear effect, into Asaoka’s equation.© CGP SITE © vO SITE | sence OF EMBANKMENT HEIGHT TD 4.0 m FOR CGH SITE SETTLEMENT {em} ° 200 400 500 ‘800 1000... 7 1200 TIME (DAYS) FIG. 16. Comparlson of Maximum Surface Settlements for Both Sites Bt (L5m,cor) 0 883 (3.0m,coP) S02 (6.0m,ccr) 384 (8.0m,coP) SSUU2.5m,PxD) 352 (5.5m, PVD) SETTLEMENT (em) TIME (DAYS) FIG. 17. Comparison of Maximum Subsurface Settlements for Both Sites The data for back analysis were obtained from field and laboratory results. Correlations of these results were also made. For the PVD ‘site, back an- alyzed parameters were applied to predict the soil behavior using o¢dometer, one-dimensional (Skempton and Bjerrum 1957; Asaoka 1978) methods. For the case of the CGP site, evaluation of settlement reductions was made by employing the equilibrium method (Aboshi et al. 1979) and design charts from the finite element formulation of Balaam et al. (1977). Evaluation ofc, or c, from Settlement Data 7 From time-settlement curves obtained from field measurements, the total settlements comprising immediate settlement p;,, primary consolidation, p,, and secondary compression, p,.. were determined. The total settlement, Pr consisting of immediate ‘and consolidation settlement, was then calculated by subtracting the effects of secondary compression. Since only conventional oedometer tests were conducted in this study, the magnitude of secondary compression was estimated based on C, = 0.04* C,, as suggested by Mesriand Castro (1987).-Employing Asaoka (1978), as suggested by: Magnan and Deroy (1980), the constant B, is related to the coefficient of consolidation as follows: For horizontal (radial) drainage only = (=D BF) (inB. oe (22) (HB) osconsstest ences 03 For vertical drainage only. a= (3) na (42) ve euneeeTE lle laseeeceeeeeeuueseueen ss (14) For the, PVD. site, Barron’s_ solution. for radial consolidation was em- ployed; for the CGP site, a:combined radial ‘and vertical flow from (9) was computed, Evaluatién of Undrained Modulus (E,) and’E,/s, Ratio The undrained modulus, E,, of the soft clay layer was evaluated from the initial ufdrainéd séttlements. Thése undrained settlements were as- sumed to be the measured settlements at the ground surfacé immediately after embankment construction. The initial settlement, ‘pj; was then ex- pressed in the following form: og . : p= az ( i ) vlecane cate huedte liege Seeetanetens cette (15) where q = the loading pressure;'B = the width of the embankment; I = an influence factor considering a rough rigid base; and ,, = the undrained modulus. Values of £,, were,then computed from (15). From the field vane test results, an average’ range of uncorrected vane shear strength, S,, was obtained and‘was used to calculate the ratio E,/S,,. Results for CGP Site The inverse analysis involves trial and error, especially at the CGP site. The complexity of three-dimensional flow, two-stage embankment loading, and the changing coefficients of consolidation are some of the factors that complicated the analyses, Using Asaoka (1978), the final settlement at the final stage of:loading was-250.0'mm..A reanalysis from this study showed that the final stress from the 2.40-m-high embankment was witliin the over- consolidation range of the underlying soil. This analysis was supported by the rapid dissipation of excess pore pressure and by observed. settlement magnitudes (Fig. 10). ‘The coefficients of consolidation tend to be very high in the overconsolidated range as a result of the clay having-very low com- pressibility for very small load increments (Bjerrum 1972). These obser- vations were proven correct from consolidation. test results and from: back calculation of settlement data from.the CGP site. Thus, in.the back ‘analysis, different c, and c, values were derived depending upon the applied load on the soil: Smaller applied load from the 2.40-m embankment height gave much higher coefficients of consolidation than the load from the.4.0-m-high embankment. These back analyzed parameters were then used in the pre- diction of the rate and amount of settlement, as shown in Fig. 18. The initial settlements were calculated, using the modified method of2-2 ASADKA (1978) METHOD OO BALAAM ot of (1877) SETTLEMENT (em) a5 ‘one -puorsou - { CEDOMETER 1 te Peet arisenican nnn ES lewrrow avo uc 157) rs ° 200 200 600) B00 1000" 1200 TWME (oars) FIG. 18. Comparison Between Predicted and Observed-Settlements for CGP Site Janbu. (Christian-and-Carrier 1978). In the abserice of data; the smeared zone of the soil surrounding the granular piles during -construction.was taken into account. by reducing to one-fifth.the constructed. final diameter of the granular piles as the effective diameter used.in the analysis. This-as; mption is in accordance-with the range tentatively recommended by Barksdale and Bacchus (1983) for the effective diameter-of granular. piles, -which varies from. one-half to one-fifteenth of the actual diameter. Asaoka’(1978) and Balaam.et al. (1977) appeared to be satisfactory in-predicting.the rate, and amount of settlements, although overprediction during the initial stages of loading was observed,-Excellent agreements.were obtained at ‘the final stage of consolidation using. back-analyzed: parameters, from this loading. ‘The equilibrium method using one-dimensional consolidation theory as well as the Skempton and Bjerrum (1957) method resulted in overestimation of the rate and amount of settlements. Results of PVD Site In the PVD site, some parameters. were assumed and tried, then compared with the actual behavior of the subsoil..A number of trials were.made based on different values of k,/k,..From these, a value of k,/k, =. 10. was found to be the most appropriate for the case under consideration with d/d,. 2.5, based on the equivalent diameter of the mandrel recommended by Hansbo (1987) and Rixner ét al. (1986). The sniéar effect was found to be equally importait, if not more, than the‘effect Of drain spacing, which is always present. This, however, also depends 6n thé chosen value of k,/k,, which has the dominant effect on the value of the back-caleulated c,,. The “wide variation in ¢, could greatly affect the-average degree of consolidation in the radial direction, U,;,and hence the prediction of the rate of settlement. This analysis was based-on the assumption that Asaoka’s method can cor- rectly predict the final settlement from one:to two years of postconstruction settlement data, or when about 60% consolidation has: already occurred. ‘The back analyzed value.of ¢, was about 4.90 n/yeare Fig. 19 shows the comparison between the-predicted and observed rates of settlement by different: methods using the back-analyzed-parameters. Except for the one-dimensional methad (without initial settlement), gen- erally all the methods predicted well the rate-and;amount of settlements.© CyI MELO) © 1.5, (LAB) FOR Cy (FIELD) © 4 cy (LAB). (© SKEMPTON AND BJERRUM (1957) 20 ONE CIMENSIONAL LOEDOMETER) Oo ASAOKA (1978) 2-6 opsenven SETTLEMENT (em) TIME (Days) FIG. 19. Comparison Between Predicted and Observed:Settlements for PVD Site Without: the inclusion of initial settlements, the one-dimensional method underestimated both the rate:and amount of'settlements at the initial stage of consolidation. Asaoka (1978) and Skempton and Bjerrum (1957) yielded very good prédictions::The one-dimensional method with the inclusion of initial settlemeiits overestimated the rate and amount of:settlements. From these analyses, it was found that c, (field)/c, (lab) ranged from 1.5— 4 for the PVD site, whilé’a wider range of 1.3-12 for-c, (field)/c, (lab) was calculated for the’ CGP site, based-on thé final stage of loading and con- solidation. The back® fanalyzed value of E, was about 3,000 kPa, giving an average E,/S, ratio of 150, where S;, is the uncorrected vane shear strength. This £,/S, ratio confirmed the earlier‘results presented by Bergado-and Khaleque (1986) and Bergado et al; (1990). Conctusions This study attempted to use inverse analysis to validate accepted theories by ‘evaluating various parameters of the soft Bangkok’ clay that had been improved by prefabricated vertical drains and compacted granular piles. On the basis of this evaluation, the following conclusions can be drawn: 1. Based on the observed preconsolidation pressures, the final settlement for the PVD site was calculated at 360.0 mm, based on Asaoka (1978). The 90% and 100% consolidation settlements occurred at 430 days and. 610 days, re- spectively, after, the beginning of construction. For the CGP site; the expected final consolidation settlement based on the observational procedure of Asaoka (1978) was 250.0:mm. This magnitude implies reduction of about 30-35% in the total final settlement using granular piles. Also, the 90% and 100% ‘con- solidations occurred at about 880 ayes and. 1,040 days, respectively, after:the beginning of construction. 2. For the PVD site, excellent agreements were attained in the predicted rate and amount of settlements using back-andlyzed parameters by the methods of Asaoka(1978)- and Skempton and. Bjerrum (1957).-At'the CGP site, good ‘agreeménts for the rate and amount of settlement resulted frorn the adoption of Asaoka (1978) and Balaanyet al. (1977), except at the initial stage of loading.3, The back-analyzed ¢, (field)/c,, (lab) for the PVD. site, ranged from-1.5— 4.0, with a back-analyzed c, of 4.90 m?/year and a back-analyzed E,/S, ratio averaging 150. For the CGP site, the range of back-calculated?c, (field)/c, (lab) was about 1.3-12, based on the final stage of loading and consolidation, 4.- For the PVD site, a value of k,/k, = 10 was found to be most appropriate with d,/d, = 2.5. For the CGP site, the effective diameter of the granular piles was taken as‘one-fifth of the constructed final diameter to account for the smeared zone. APPENDIX I. REFERENCES Aboshi, H., Ichimoto, E., Enoki, M., and Harala, K. (1979). “The cainporer= 4 method to improve characteristics of soft clays by inclusion of large diameter sand columns.” Proc. Ist Int. Conf. on Soil Reinforcement, Paris, 1, 211-216. Asaoka, A. (1978). “Observational procedure for settlement prediction.” Soils Found., 18(4), 87-101. Balaam, N. P., Poulos, H. G., and Brown, P. T. (1977). “Settlement analysis of soft clays reinforced with granular piles.” Proc. 5th Sowihwest Asian Conf. on Soil Engrg., Southeast Asian Geotechnical Society, Bangkok, Thailand, 81-92. Barksdale, R. D., and Bachus, R..C. (1983). ‘Design and construction of stone columns.” Report No. FHWAIRD-83/026, Nat. Tech. Info. Service, Springfield, Va. Barron, R. A. (1948). “Consolidation of fine-grained soils by drain wells.” Trans. ASCE, 124, 709-739. Bergado, D. T., and Khaleque, M.A. (1986). “Correlation of LLT pressuremeter, vane, and Dutch cone tests in Bangkok marine clay, Thailand.” The Pressuremeter and Its Marine Applications: Second Int. Symp., J. L. Briand and J. M. E. Audibert, eds., STP 950, ASTM, Philadelphia, Pa., 339-353. Bergado, D. T., Miura, N., Singh, N., and Panichayatum, B. (1988a). “Improvement of soft ‘Bangkok clay’ using vertical band drains based on full scale test.” Proc. Int. Conf. on Engrg. Problems of Regional Soils, Pergamon Press, Beijing, China, 379— 384. Bergado, D. T., Miura, N., Panichayatum, a and. Sampaco, C. L..(1988b). “Re- inforcement of soft Bangkok clay using granular piles.” Proc. Int. Geotech. Symp. on Theory and Practice of Earth Reinforcement, Balkema, Rotterdam, TheNeth- erlands, 179-184. Bergado, D. T., Ahmed, S., Sampaco, C. L., and Balasubramaniam, A. S. (1990). “Settlement of Bangna-Bangpakong highway on soft Bangkok clay.” J. Geotech. Engrg., ASCE, 116(1),:136—155. Bjerrum, L. (1972). “Embankments on soft ground.” Proc. Spec. Conf. on Per- formance of Earth and Earth-Supported Structures, ASCE, New York,.N.Y., 2, 1-54, Carillo, N. (1942). “Simple two- and three-dimensional casés in the theory of con- solidation of soils.” J. Math. Phys., 21(1), 1-5. Christian, J. T., and Carrier, W. D., IU. (1978). “Janbu, Bjerrum and Kjaernsli’s chart reinterpreted.” Can. Geotech. J., 15, 123-128. Hansbo, S. (1979). “Consolidation of clay by band shaped:prcfabricated drains.” Ground Engrg., 12(5), 16-25. Hansbo, S. (1987). “Design aspects of vertical drains and lime column installations.” Proc. 9th Southeast Asian Geotech. Conf., Southeast Asian Geotechnical Society, Bangkok, Thailand, 2, 8-1 to 8-12. Juran, L., and Guermazi, A. (1988). “Settlement response of soft soils reinforced by compacted sand columns.” J. Geotech. Engrg., ASCE, 114(8), 930-943. Laminen, K., and Rathmeyer,’H. (1983).-“Experience on horizontal oedometer test to determine the parameter c,.”. Prac. 8th European Conf..on Soil Mech. and Found. Engrg., Balkema, Rotterdam, The Netherlands, 2,-647-651. Magnan, J. P., and Deroy, J. M. (1980). “Analyse graphic des courbes de consol-(Mesri, idation oedometrique.” Bull. Liason Lab. Ponts et Chaussees, Paris, France, 9, sand:Castro, A. (1987). “C,/C, concept and K, during secondary compres- sion.” J. Geotech. Engrg., ASCE, 113(2),:230-247. Newmark, N. (1942). “Influence charts for computations'of stresses in elastic’ foun- dations.” Bull. No, 338, Univ, of Illinois, Urbana, It. “Olson, R. E. (1977). “Consolidation under time-dependent loading,” Jy Geotech. » Engrg. Div., ASCE, 103, 55-61. Poulos, H. G., and Davis, E. H. (1974). “Elastic solutions for soil and rock me- chanics.” John Wiley & Sons, Inc., New York, N.Y. Rixner, J. J., Kraemer, S. R., and Smith, A. D. (1986). “Prefabricated vertical drains, Vol. 1: Engineering guidelines.” Report No. FHWA/RD-86/168, Nat. Tech. Info. Service, Springfield, Va. Skempton, A. W., and Bjerrum, L. (1957). “A contribution to the settlement analysis of foundations on clay.” Geolechnique, 7,.168-178. ApPeENDIX Il. “NOTATION * The following symbols are used in this paper: thickness of band-shaped drains; — 7 * area replacement ratio; width of embankment; width of band-shaped drains; coefficient of secondary compression; horizontal coefficient of consolidation; vertical coefficient of consolidation; diameter of equivalent soil cylinder; diameter of smeared zone; equivalent diameter of drain; F(n) + F, + Fy factor to account for well resistance; factor to account for smear; * factor to account for drain spacing; thickness of treated ground; influence factor; permeability along horizontal direction; permeability at smeared zone; drain’ length; stress concentration factor; loading intensity; discharge capacity of drain; radial distance; settlement of composite-ground; combined average degree of consolidation due to vertical and horizontal drainage; averagé degree of consolidation due to horizontal drainage; average degree of consolidation due to vertical drainage; excess, pore pressure due to. radial flow; z= distance-from open end of drain; SBi slope of settlement records; ‘time interval; v = undrained Poisson ratio;drained Poisson ratio; primary consolidation settlement; immediate settlement; settlement at time #,; net total settlement; secondary settlement; average loading intensity; '¢ = Stress intensity in cohesive soil; and o, = Stress intensity in granular piles.